Merge branch 'main' into fix_post_comment_modlog_filtering

crates/api_common/src/request.rs

@@ -105,7 +105,11 @@ pub fn generate_post_link_metadata(
     }
     // Generate local thumbnail if allowed
     else if allow_generate_thumbnail {
-      match post.url.or(metadata.opengraph_data.image) {
+      match post
+        .url
+        .filter(|_| is_image_post)
+        .or(metadata.opengraph_data.image)
+      {
         Some(url) => generate_pictrs_thumbnail(&url, &context).await.ok(),
         None => None,
       }

crates/api_common/src/utils.rs

@@ -536,25 +536,8 @@ pub async fn get_url_blocklist(context: &LemmyContext) -> LemmyResult<RegexSet>
       .try_get_with::<_, LemmyError>((), async {
         let urls = LocalSiteUrlBlocklist::get_all(&mut context.pool()).await?;
 
-        let regexes = urls.iter().map(|url| {
+        // The urls are already validated on saving, so just escape them.
-          let url = &url.url;
+        let regexes = urls.iter().map(|url| escape(&url.url));
-          let parsed = Url::parse(url).expect("Coundln't parse URL.");
-          if url.ends_with('/') {
-            format!(
-              "({}://)?{}{}?",
-              parsed.scheme(),
-              escape(parsed.domain().expect("No domain.")),
-              escape(parsed.path())
-            )
-          } else {
-            format!(
-              "({}://)?{}{}",
-              parsed.scheme(),
-              escape(parsed.domain().expect("No domain.")),
-              escape(parsed.path())
-            )
-          }
-        });
 
         let set = RegexSet::new(regexes)?;
         Ok(set)

crates/utils/src/utils/validation.rs

@@ -309,21 +309,44 @@ pub fn is_url_blocked(url: &Option<Url>, blocklist: &RegexSet) -> LemmyResult<()
   Ok(())
 }
 
+/// Check that urls are valid, and also remove the scheme, and uniques
 pub fn check_urls_are_valid(urls: &Vec<String>) -> LemmyResult<Vec<String>> {
   let mut parsed_urls = vec![];
   for url in urls {
-    let url = Url::parse(url).or_else(|e| {
+    parsed_urls.push(build_url_str_without_scheme(url)?);
+  }
+
+  let unique_urls = parsed_urls.into_iter().unique().collect();
+  Ok(unique_urls)
+}
+
+pub fn build_url_str_without_scheme(url_str: &str) -> LemmyResult<String> {
+  // Parse and check for errors
+  let mut url = Url::parse(url_str).or_else(|e| {
     if e == ParseError::RelativeUrlWithoutBase {
-        Url::parse(&format!("https://{url}"))
+      Url::parse(&format!("http://{url_str}"))
     } else {
       Err(e)
     }
   })?;
 
-    parsed_urls.push(url.to_string());
+  // Set the scheme to http, then remove the http:// part
+  url
+    .set_scheme("http")
+    .map_err(|_| LemmyErrorType::InvalidUrl)?;
+
+  let mut out = url
+    .to_string()
+    .get(7..)
+    .ok_or(LemmyErrorType::InvalidUrl)?
+    .to_string();
+
+  // Remove trailing / if necessary
+  if out.ends_with('/') {
+    out.pop();
   }
 
-  Ok(parsed_urls)
+  Ok(out)
 }
 
 #[cfg(test)]
@@ -600,17 +623,21 @@ mod tests {
 
   #[test]
   fn test_url_parsed() {
+    // Make sure the scheme is removed, and uniques also
     assert_eq!(
-      vec![String::from("https://example.com/")],
+      &check_urls_are_valid(&vec![
-      check_urls_are_valid(&vec![String::from("example.com")]).unwrap()
+        "example.com".to_string(),
+        "http://example.com".to_string(),
+        "https://example.com".to_string(),
+        "https://example.com/test?q=test2&q2=test3#test4".to_string(),
+      ])
+      .unwrap(),
+      &vec![
+        "example.com".to_string(),
+        "example.com/test?q=test2&q2=test3#test4".to_string()
+      ],
     );
 
-    assert!(check_urls_are_valid(&vec![
+    assert!(check_urls_are_valid(&vec!["https://example .com".to_string()]).is_err());
-      String::from("example.com"),
-      String::from("https://example.blog")
-    ])
-    .is_ok());
-
-    assert!(check_urls_are_valid(&vec![String::from("https://example .com"),]).is_err());
   }
 }